Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    attends to tokens formatted with small caps from tokens formatted with square brackets

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.06
    1:0.08
    2:0.09
    3:0.09
    4:0.06
    5:0.06
    6:0.39
    7:0.14
    Negative Logits
    Vidite
    -0.29
     maat
    -0.29
    Minn
    -0.28
     isolating
    -0.28
    декс
    -0.27
    nya
    -0.27
     charts
    -0.26
     be
    -0.26
     Louisiana
    -0.26
     Becher
    -0.26
    POSITIVE LOGITS
     initComponents
    0.47
    routeProvider
    0.45
     defaultstate
    0.44
    ContentAsync
    0.43
    enumii
    0.43
     AttributeSet
    0.42
    ftagPool
    0.42
    ."</
    0.41
     }}"></
    0.41
    awtextra
    0.40
    Act Density 0.001%

    No Known Activations